vi capsular polysaccharide-protein conjugates · pdf filevi capsular polysaccharide-protein...

Vi CAPSULAR POLYSACCHARIDE-PROTEIN CONJUGATESFOR PREVENTION OF TYPHOID FEVER

Preparation, Characterization, and Immunogenicity inLaboratory Animals

BY SHOUSUN CHEN SZU, AUDREY L. STONE, JOAN D. ROBBINS,RACHEL SCHNEERSON, AND JOHN B. ROBBINS

From the Laboratory ofDevelopmental and Molecular Immunity, National Institute ofChildHealth and Human Development, National Institutes ofHealth, Bethesda, Maryland 20892;and the Division ofBiochemistry and Biophysics, Office ofBiologics, Research and Review, Food

and Drug Administration, Bethesda, Maryland 20892

Enteric fevers continue to cause considerable morbidity and mortality innations that have not yet achieved control of drinking water and sewage disposal(1, 2) . In these countries, the most frequent and serious cause of enteric feversis Salmonella typhi (typhoid fever) (2). Immunoprophylaxis against typhoid feveron a world-wide basis has not been attempted because the two available vaccineshave limitations ; (a) cellular typhoid vaccines induce only a limited immunityand elicit side reactions that are frequent and severe enough to have discouragedtheir widespread acceptance (2, 3) ; and (b) an orally administered attenuatedstrain of S. typhi, Ty-21a, requires three to four doses to induce about 65%protection, it is expensive, and its mode of protection has not been identified,which has prevented precise standardization of the vaccine (1, 4) .

Recently, two clinical evaluations in populations with high rates of typhoidfever (- ..1 %/annum) have provided evidence that immunization with the capsularpolysaccharide (CPS)' of S. typhi (Vi) confers immunity against typhoid fever (5,6) . The Vi vaccine, prepared under conditions that did not change its structure,elicited a fourfold or greater rise in serum antibodies in -75% of children andadults in Nepal and in school children in the Eastern Transvaal, Republic ofSouth Africa (2, 7, 8) . The protective efficacy of the Vi in these two trials was^-70% . In contrast, the same Vi elicited a _>4-fold antibody rise in 97% of youngadults in France and the United States (9). The seroconversion rate and efficacyof other CPS, e.g ., meningococcal vaccines, were also lower in Africa than inFinland or the United States (10-12). This lesser immunogenicity and efficacyof meningococcal vaccines was attributed to the high burden of infections,This work was supported in part by a contract, Project Order MIPR 85 PP5814, from the WalterReed Army Institute of Research . Address correspondence to S. C. Szu, NIH, Bldg . 6, Rm A1-10,Bethesda, MD 20892.

' Abbreviations used in this paper:

Vi, Vi capsular polysaccharide ; CPS, capsular polysaccharide ;SPDP, N-succinimidyl 3-(-2-pyridyldithio) propionate ; EDAC, 1-ethyl-3-(3-dimethylaminopropyl) car-bodiimide; TT, tetanus toxoid ; DT, diphtheria toxoid ; CT, cholera toxin ; RIA, radioimmunoassay ;NMR, nuclear magnetic resonance ; FTIR, Fourier-transformed infrared spectroscopy .

1510

Journal of Experimental Medicine - Volume 166

November 1987

1510-1524

SZU ET AL .

1511

including malaria, in the African population. Since the protective responseelicited by CPS vaccines is serum antibodies, it could be predicted that a moreimmunogenic Vi would be more protective against typhoid fever in high-riskpopulations .

Originally, Goebel and Avery (13, 14) showed that the immunogenicity ofpneumococcus type 3 polysaccharide could be increased by binding it chemicallyto a carrier protein . This principle has been applied successfully to increase theimmunogenicity of CPS of other pathogens (15-23) . We developed methods forsynthesizing covalent bonds between the Vi and proteins to both increase theimmunogenicity of and to confer the property of T-dependence to this antigen .Vi, a linear homopolymer of -4-D-a-NAcGalA-(1- is O-acetylated up to 90% atC3 (24) . The scheme used to prepare these conjugates used the carboxyl functionof the NacGalA to form a thiol derivative . This thiol derivative was combinedwith proteins derivatized by the thiol-active compound, N-succinimidyl-3-(2-pyridyldithio) propionate (SPDP) (25) . The standardization and immunologicproperties of several model Vi-protein conjugates formed by this method arereported here .

Materials and MethodsVi CPS (Vi) .

The Vi used for the conjugates was prepared from Citrobacterfreundii,strain WR 7011, kindly given to us by Dr . Lewis Baron, Walter Reed Army Institute ofResearch, Washington, DC. C . freundii was cultivated in medium containing yeast extractdialysate as described (3) . Vi was precipitated from the culture supernatant with 1%hexadecyltrimethyl ammonium bromide (Sigma Chemical Co., St . Louis, MO) and se-quentially treated with DNase, RNase, and then pronase (19, 27, 28) . The enzyme-treatedproduct was purified with cold phenol and the LPS was removed by centrifugation at60,000 g, 10 0'C, for 5 h (10). The Vi was dialyzed exhaustively against pyrogen-free waterand freeze-dried . Vi from S. typhi was prepared by a modification of this method by theInstitut Merieux, Lyon, France (9) . The final products contained <1% protein or nucleicacid (26) and <0.01 % LPS as measured by SDS-PAGE (29) . The molecular size of the Vipreparations was heterogeneous ; the main peak had a molecular mass 5 x 10 s kD asestimated by gel filtration through Sephacryl S-1000 (Pharmacia Fine Chemicals, Pisca-taway, NJ) equilibrated in 0.2 M NaCl . A lower molecular size Vi, ^- 65 kD was preparedby ultrasonic irradiation (30) .

Proteins .

BSA (Sigma Chemical Co.) and cholera toxin (CT, Lot 582 ; Institut Merieux)were used without further purification . Tetanus toxoid (TT ; Institut Merieux) anddiphtheria toxoid (DT ; Rijks Instituut voor Volksgezonheid, Bilthoven, Netherlands) werefurther purified by gel filtration through Sephacryl S-300 (Pharmacia Fine Chemicals)equilibrated in 0.2 M NaCl (17, 22) . The fractions, corresponding to the molecular weightof the two toxoids, were concentrated by ultrafiltration and passed through 0.45-/mfilters (Millipore Co., Bedford, MA). CT was additionally characterized for toxicity by theChinese hamster ovary (CHO) cell assay and by the intradermal rabbit skin test (kindlyperformed by Dr . John Craig, State University of New York, Brooklyn, NY) (31) .

Direct Binding ofVi to Proteins with EDAC (Vi-DT,).

Vi was bound to DT by the methodof Beuvery et al . (16) . Equal volumes of Vi and DT, containing 10 mg/ml each, weremixed, the pH was adjusted to 5 .0 with 0.1 N HCI, and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDAC) (Bio-Rad Laboratories, Richmond, CA) was added to a final con-centration of 0.1 M . The pH was maintained at 5.0 by addition of 0.1 N HCI for 3 h atroom temperature . The reaction mixture was stirred at 3-8°C for an additional 24 h,dialyzed against 0.2 M NaCl for 48 h, and then centrifuged at 10,000 g, for 2 h, at 3-8°C . The supernatant fluid was subjected to gel filtration through 4B-Sepharose equili-brated in 0.2 M NaCl . The fractions were monitored by refractometry and by absorption

1512

Vi POLYSACCHARIDE-PROTEIN CONJUGATES

at 2,800 A . The void-volume fractions, which contained the Vi-DT, conjugate, werepooled, made 0.01 % in thimerosal (Eli Lilly and Co., Indianapolis, IN), and stored at 3-8°C.

Thiolation of Vi with Cystamine.

Vi, 10 mg/ml in 0 .2 M NaCl, was allowed to dissolveovernight at 3-8°C . Cystamine or cysteamine (Calbiochem-Behring Corp., La Jolla, CA)at twice the weight of the Vi was then added as a powder to the Vi solution . The pH ofthe reaction mixture was adjusted to 4.9 with 0.1 N HCl and the temperature wasmaintained at 37° C . EDAC, in an amount equal to the weight of the Vi, was added tothe reaction mixture with stirring. The pH was maintained at about 4.9 by the additionof 0.1 N HCl until stable . The reaction mixture was then dialyzed exhaustively againstdistilled water at 3-8'C and freeze-dried. The extent of thiolation was estimated by theiodoplatinate assay (32) using cystamine as the reference . The thiol ester content wasquantitated after reduction of the Vi derivative with DTT (Sigma Chemical Co.) andpassage of the reaction mixture through Bio-Gel P10 (Bio-Rad Laboratories) to removethe low molecular weight materials . The material eluted in the void volume was thenassayed for its SH content using cysteamine as a standard (33) .

Derivatization ofProteins with SPDP.

Protein solutions were made within a range of 5-20 mg/ml . Free SH groups were blocked by treatment with 0.01 M iodoacetic acid (SigmaChemical Co.) . The reaction mixture, containing the iodoacetic acid and the protein, wasincubated at room temperature for 1 h and then equilibrated against 0.15 M Hepes(Boehringer Mannheim Biochemicals, Indianapolis, IN), 5 mM EDTA, pH 7.55 (HEbuffer) by dialysis overnight at 3-8°C . SPDP (Pierce Chemical Co., Rockford, IL), 20mM in 99.5% ethanol, was added to the protein solution with stirring (25) . The finalmolar ratio of SPDP to protein ranged from 5-25 depending upon the extent ofderivatization desired . The reaction was allowed to proceed for 1 h at room temperature .The excess reagent was removed by dialysis against HE buffer overnight at 3-8 °C followedby gel filtration through Bio-Gel P10 in HE buffer. The void volume was pooled,concentrated to 5-10 mg/ml, and stored at 3-8'C . The molar ratios of 2-pyridyldisulphide in the derivatized protein were determined by reducing the disulfide bond in50 mM DTT and calculating the amount of pyridine-2-thione released by using theextinction coefficient for pyridine of 8 .08 x 10 s/mot/cm at 3,430 A (23, 25) .

Synthesis of Vi-Protein Conjugates.

The cystamine derivative of Vi was dissolved in HEbuffer, at 5-10 mg/ml. DTT was added to a final concentration of 100 mM and stirredfor I h at room temperature . After dialysis against HE buffer for 2 h, the reaction mixturewas then passed through a Bio-Gel P10 column, equilibrated with HE buffer and the voidvolume peak was concentrated by ultrafiltration (Amicon Corp., Danvers, MA; molecularweight cutoff >5,000) under N2 pressure . SPDP-derivatized protein was added to thereduced Vi derivative to achieve an equimolar ratio ofN-pyridyl disulfide groups and SHgroups . The reaction mixture was flushed with N2 and allowed to incubate at roomtemperature for 1 h, and then at 3-8°C for 24 h . The release of pyridine during thereaction was measured by the change in OD 3,430 A . The reaction mixture was concen-trated by ultrafiltration (see above) and subjected to gel filtration through Sephacryl 5-1000 in 0.2 M NaCl at room temperature . The void volume fractions were pooled,concentrated by ultrafiltration, dialyzed against 0.15 M NaCl, 0.01 % thimerosal (Eli Lilyand Co .), I mM EDTA, pH 7.0, and stored at 3-8 °C . The protein concentration of theconjugate was determined by the Coomassie Blue assay (Bio-Rad Laboratories) using theSPDP-derivatized protein as the standard (34) . The concentration of protein and Vi inthe conjugates were also determined by Fourier transformed infrared spectroscopy (FTIR)and by spectrophotometric titration by acridine orange (35) .

Pneumococcus Type 6B-TT Conjugate (Pn6B-TT).

Pneumococcus type 6B CPS, (InstitutMerieux), was derivatized with adipic acid dihydrazide and bound to TT as described (17,20, 22) . The protein/polysaccharide ratio of this preparation was 2.8 .

Immunization of Mice .

Female, weanling, BALB/c mice, 16-20 g, were injected sub-cutaneously with 0.1 ml of either Vi, Vi conjugates or saline 1, 2, or 3 times at 2 wk apart .Mice from each experimental group were exsanguinated 10 d after each injection . Alum-adsorbed Vi-CTX� was prepared with Alhydrogel (Superfos; Kemi A/S, Copenhagen,

SZU ET AL .

151 3

Denmark) . The Alhydrogel was centrifuged and Vi-CTxu was added to the pellet toachieve a final concentration of 0.5 mg aluminium/ml and 5 .0 Kg of Vi as a suspension .This mixture was tumbled overnight at room temperature and then stored at 3-8°C .

Immunization ofPrimates .

Juvenile Rhesus monkeys, housed at the Division of ProductQuality Control, Office of Biologics Research and Review, FDA, were injected subcuta-neously twice, 1-mo apart with 0.5 ml of Vi or Vi-CTx, i containing 25 kg of Vi . Controlswere injected with Pn6B-TT containing 15 ttg of CPS (20) . The monkeys were bledbefore and 3 wk after each injection .

Immunological Methods.

Serum Vi antibodies were measured by RIA and expressed as,ug antibody/ml (9) . The differences between the levels of antibodies in experimentalgroups were calculated by Fisher's exact t-test . The results were tabulated as the geometricmean and 80% confidence limits . Antibodies to CT were measured by ELISA (36) . Rabbitanti-BSA serum was obtained from Cappel Laboratories, Cochranville, PA. Hyperimmuneburro antiserum, (B-260) containing 660 Ag Vi antibody/ml, was prepared by multipleintravenous injections of formalin-fixed S . typhi Ty2 as described (37) . The preparationand characterization of burro 241, hyperimmune CT antiserum, containing 16 .5 mgantibody/ml, has been reported (38) . Rocket immunoelectrophoresis and immunodiffu-sion were performed as described (23) .FTIR Spectroscopy .

The composition of the conjugates was determined using the Viand proteins as references . 1 mg of Vi, carrier protein, or Vi conjugate was added to 100mg of KBr, and dissolved in 2.0 ml distilled water . The sample was freeze-dried andpressed into a pellet . FTIR spectra were recorded on a Nicolet 7199 spectrometer andanalyzed as described (35) .f "CJNuclear Magnetic Resonance (NMR).

[' gC]NMR spectra ofthe Vi (20 mg/ml D20)were recorded at 60°C in a Nicolet 270 spectrometer . A 5 .0-mm sample cell was usedand the spectrometer was operated at 67.9 MHz in the pulse Fourier-transformed modewith complete proton decoupling and quadrature phase detection . A 4-Hz line broadeningwas applied to the signal before Fourier transformation to enhance the signal-to-noiseratio .SDS-PAGE. The molecular weights of the proteins and their Vi conjugates were

assayed in 7.5% polyacrylamide gels (23). Samples containing 5-10 ug of protein with orwithout 2-ME were electrophoresed concurrently with protein standards (Pharmacia FineChemicals) . The gels were stained with Coomassie Blue (Sigma Chemical Co.) .

ResultsVi Antigens .

The Vi from S . typhi and some strains of C . freundii is a linearhomopolymer of NacGaIA acid, variably O-acetylated at the C3 position (24) .FTIR and [ 1sC]NMR spectroscopy showed that the structures of the Vi from theWR 7011 strain of C.freundii and the Ty2 strain ofS. typhi were indistinguishable(39, 40) (Fig. 1) . Immunodiffusion analyses showed a reaction of identity betweenVi polysaccharides from C. freundii and S. typhi when reacted with B-260 anti-Viantiserum (data not shown) . The Vi used for immunization and for synthesis ofthe conjugates in these experiments contained about 80% 0-acetyl per molerepeating unit as determined by [' 3C]NMR (Fig . 2) . Since the Vi has no vicinalhydroxyl groups, the carboxyl served as an activation site in the strategy for theconjugation reaction .

Direct Binding of Vi with Proteins Using EDAC.

The first Vi conjugates wereprepared with EDAC, which catalyzed the direct conjugation between thecarboxyls of Vi and the amino groups of DT (16). The yield from this reaction,illustrated by a representative product Vi-DTI , was ^-3% ofthe starting materials .Vi antibody levels elicited in mice by this conjugate and by the two controls, Viand saline, are given in Table I .

1514

a


}3OAc

- 4)-a-D-Gal NAc A-(1 --

b

cHIS -(CH 2 )= -NH 2

2HN-(CH2) 2 -S-S-(CH 2) 2 -NH Zcysteamine cystamine

TABLE I

Serum Antibody Responses ofFemale BALE/c Mice Injected with Vi or Vi-DT Conjugate (Vi-DTI) Prepared with EDAC as the coupling reagent

FIGURE 1 .

(a) Structure of the repeating unit ofthe Vi capsular polysaccharide of Salmonella typhi.(b) Structure of cysteamine . (c) Structure of cys-tamine.

r0 PPM

FIGURE 2. ['"CJNMR chemicalshift of Vi depolymerized by ul-trasonic irradiation (65,000 molwt, 20 mg/ml concentration),downfield from "C-enriched ac-etonitrile used as an internal stan-dard . The spectra were recordedfor a D20solution in a 5-mm tubewith a Nicolet 270 spectrometeroperating at 67.89 MHz in pulseFourier-transformed mode .

Mice were injected subcutaneously 2 wk apart with 0.1 ml of each vaccine or saline . There wereeight mice for each experimental group. 10 d after each injection mice were exsanguinated andtheir sera were assayed for Vi antibodies by RIA (9) . Results are expressed as the geometric meanand 80% confidence limits .vs . ** : p = 0.0014, ** vs . tt, III : P = 0.001, o vs . "I, $ vs . H : P = 0.001, 1 vs .

vs. ***: P = 0.0001,# vs, ff : P = 0.003 .

There were no Vi antibodies in preimmune sera or from mice injected withsaline (controls) in this and subsequent experiments. Dosages of 0.5, 5.0, or 50.0ug of Vi elicited similar levels of antibodies after the first, second, and thirdinjections ; the slightly higher levels of antibodies in the group injected with 5.0

Ag of Vi after the third injection were not statistically different. The levels ofantibodies after the second and third injections of 5.0 or 50.0 Ag of Vi-DTI were

Vaccine Dose1st

Vi antibody

2nd 3rd

wg uglml ug/ml Ag/mlVi 0.5 0.19* (0.05-0.72) 0.61 (0 .30-1 .23) 0.47 (0 .19-1 .15)

5.0 0.56 (0 .28-1 .11) 0.431 (0 .13-2.33) 1 .11 (0.29-4.26)50.0 0.30 (0.13-0.75) 0.24 11 (0 .04-1 .47) 0.20" (0.04-0.97)

Vi-DT, 0.5 0.07** (0.04-1 .12) 0.31 (0 .06-1 .52) 0.79 (0.28-2.24)5 .0 0.63*$ (0 .30-1 .30) 2.790 (0.65-12 .0) 2.85ff (1 .09-7.46)

50.0 1 .4011 (0 .80-2.45) 2.65" (0.69-10 .1) 3.00*** (1.39-6.46)

SZU ET AL.

151 5

0N

3

FIGURE 3 .

Gel filtration profileof Vi-CTx � and its components,Vi and CT, through S1000 Se-phacryl equilibrated in 0 .2 MNaCl . (A) Conjugate . (0) refrac-tive index, (0) OD 2,800 t\ . (B)Vi-cystamine . (*) refractive in-dex, and CT, (A) OD 2,800 t1 .

about sevenfold higher than those elicited by the Vi alone (p = 0.001) . Each ofthe three dosages of Vi-DTI elicited a booster response after the second injection .No differences in the levels of antibodies elicited by the 5.0- and 50.0-tag dosesof Vi-DT, were observed; both of these doses were more immunogenic than 0.5tig of this conjugate after the first injection only (p = 0 .001) . Based upon thesedata, the reports of Gaines et al . (41) and Landy (42), and our experience withother conjugates (17, 19), all mice were immunized with 2 .5 ug of the Vi, aloneor as a component of a conjugate, in ensuing experiments .

Vi-Protein Conjugates with SPDP .

An alternate conjugation procedure, whichcovalently bound thiolated derivatives of the Vi- and the SPDP-derivativedproteins, was studied in order to increase the yield and immunogenicity of theVi conjugates.

Introduction of thiol groups onto the Vi was attempted by forming amidebonds between the amino groups of cysteamine (Fig . 1 b) and the carboxyls ofthe Vi in the presence of EDAC. The yield of thiol added to the Vi by thisscheme was <1% wt/wt, probably because the SH groups of cysteamine wereoxidized . To avoid this possibility, cystamine was used, the thiols of which arelinked by disulphide bonds and are thus protected from oxidation (Fig . 1 c) .Using EDAC to catalyze amide bond formation, ^-6% wt/wt of cystamine wasbound to the Vi . After reduction of the disulphide bonds by DTT, the yield ofthiol esters was 0 .5-2% of the repeating monosaccharide . The freeze-driedcystamine derivative of the Vi was stable at -20°C.

Physicochemical Characterization of the Vi Conjugates.

The disulfide bonds ofthe Vi-cystamine derivative were reduced with DTT before the conjugationreaction . The molecular size of a representative conjugate, Vi-CTvJII (<O.1 kD)was larger than that of the proteins (0 .64 kD) or the Vi (0.36 kD), as illustratedby the gel filtration profile (Fig . 3) . The composition of the thiolated Vi, theSPDP-protein derivatives, and the conjugates used in these studies are listed inTable II . The change in the molecular size of the intermediate and final productsof the conjugation reaction was further analyzed by SDS-PAGE (Fig . 4) . TheBSA-SPDP derivative exhibited a pattern similar to that of native BSA, demon-strating that aggregation did not occur during the reaction with SPDP. The Vi-BSA,v , as did all the Vi conjugates, failed to enter the gel, probably due to its

1516


TABLE II

Characterization of Vi-Protein Conjugates Prepared with N-Succinimidyl 3-(2-pyridylthio)Propionate (SPDP)

* Percent of Vi recovered in conjugate.

FIGURE 4. Polyacrylamide gel (7.5%) electrophoresispattern of BSA and its conjugate with Vi . (Lanes 1-4)The migration pattern of samples reduced by 2-ME, (5to 8) samples not treated by 2-ME . (2 and 5) Vi-BSA,v,(3 and 6) BSA-SPDP, (4 and 7) BSA, (1 and 8) proteinstandards.

FIGURE 5. Immunodiffusion pattern of Vi-BSA,v conju-gate reacting with various antisera in 1 % agarose gel. Thecenter well contained 15 Ml of Vi-BSA,v (0 .5 mg/ml) . Wells1 and 4, Burro 260 anti-S . typhi Ty2 antiserum; wells 2 and3, rabbit anti-BSA antiserum.

large size . Vi-BSA,v, reduced with 2-ME, exhibited a band similar to the nativeBSA, indicating that the Vi was bound to this protein by a disulphide bond .

Antigenic Analysis .

The Vi conjugates were analyzed by immunodiffusion withantisera to each of its components (Fig. 5) . Theanti-Vi and anti-BSA sera formed

Conjugate Yield* SH/Vi SPDP/protein Protein/Vi

wt/wt mol/mol wt/wt wt/wt

Vi-BSA,v 63.0 1 .2 17 .0 0.078 0.7Vi-CTv � , 5.4 1 .0 5.4 0.019 0.4Vi-CT,� 18.0 0.5 4.5 0.016 1 .5Vi-DTxv 14.9 0.5 5.0 0.024 1 .6Vi-DTxx 16.8 0.4 3.8 0.018 1 .5Vi-TTXXv 6.2 0.7 12 .0 0.025 1 .4

SZU ET AL.

1517

TABLE III

Serum Antibody Responses ofMice Injected with the Vi or Vi Conjugates with BSA, CT, DT orTT Crosslinked with SPDP

Female, I6-20 g, BALB/c mice were injected subcutaneously with 0.2 ml containing 2.5 Ag Vior Vi conjugate 2 wk apart . Mice of each group were exsanguinated 10 d after their last injectionand Vi antibodies were measured by RIA (59) .

* Numbers indicate geometric mean (80% confidence limits).t vs . 11 ,', *** M, of, III : p = 0.001 ; $ vs. 1 : p = 0.002 ; t vs . II : p = 0.04; ** vs.' : p = 0.0003; # vs .t# : p = 0.003 .Adsorbed onto aluminium hydroxide, 0.5 mg A1 2"/ml .

" Molecular size of Vi reduced to ^x65,000 by ultrasonic irradiation (30) .

a partial identity reaction with the conjugate indicating that the Vi and Vi-BSAwere covalently bound. The Vi antiserum reacted with the Vi and Vi-BSA,v witha partial identity reaction ; a spur of precipitation overrode the Vi line on to thereaction of the antiserum with the Vi-BSAiv (not shown). Conjugates preparedwith DT, TT, and CT failed to react with their homologous antiprotein antiseraby immunodiffusion but did precipitate with these antisera by rocket immunoe-lectrophoresis (not shown) .

Residual Toxicity of Vi-CTx� Conjugate.

The toxicity of CT toxin was reducedby its conjugation to Vi . A 10'-fold reduction was observed in the CHO cellassay . The skin test in rabbits however, showed only a 1O'-fold reduction intoxicity . Although injection with 0.5 ml of Vi-CTxct (one human dose) had noeffect on mice, injection of 5.0 ml i.p . of Vi-CTx� resulted in the death of twoof three guinea pigs . This conjugate, therefore, did not pass the safety require-ments of the Code of Federal Regulations (CFR 600:16) .

Immunogenicity in Laboratory Mice .

None of the preimmunization sera or serafrom mice injected with saline had detectable levels of Vi antibodies (Table 111) .All of the conjugates elicited higher levels of antibodies than the Vi alone (p =0.001) . The highest level of antibodies after the first injection were elicited bythe Vi-CT prepared from the Vi depolymerized by ultrasonic irradiation, butthese differences were not statistically significant . Booster responses after thesecond injection, as defined by a fourfold or greater increase in the geometricmean antibody levels, were observed in the animals injected with all the conju-gates except the Vi-CT adsorbed . Vi-TTxxv also elicited an increase in antibodiesafter the third injection . The levels in this group however, were similar to thoseobserved with the other conjugates after two injections . The poorest response

Immunogen n1st injection

Antibody*

2nd injection 3rd injectionPg/ml Wg/ml pg/ml

Vi 7 0.56$ (0.22-1 .43) ND NDVi-BSA,v 7 2.861 (1 .25-6.53) ND NDVi-CTv�, 4 1 .351 (0 .27-6.71) 4.53 (2.74-7 .46) NDVi-CTx� 10 2.12' (1 .29-3.47) 9.04** (4.84-16 .9) 6.28# (1.98-19 .9)Vi-CTxufl 10 1 .0111 (0 .57-1 .79) 1.16 (0.54-2.53) 0.87 (0.37-2.07)Vi-CT UD" 10 2.94*** (1 .55-5.55) 4.23 (1 .75-10 .2) NDVi-DTxv 10 2.57#$ (1 .27-5.17) 4.61 (2.49-8.52) NDVi-DTxx 10 1 .76 (0.92-3.38) 1.26 (0.63-2.52) NDVi-TTxxv 9 2.56111 (1 .40-4.69) 3 .15 (1 .71-5.80) 4.18 (2.74-6.36)

151 8


TABLE IVSerum Antibody Responses ofJuvenile Rhesus Monkeys Immunized with Vi, Vi Conjugated to

CT (Vi-CTx�), or Pneumococcus Polysaccharide type 6B Conjugated to TT (Pn6B-TT)

Discussion

* Geometric mean (80% confidence limits) . Measured by RIA (9). 1 vs.' : p = 0.004,'vs . ** : 0.02, N

vs . ** : p = 0.0003,' vs . t: p = 0.009, ** vs . 1 :p=0.001 . Monkeys were immunized subcutaneouslywith 0.5 ml containing either 25 ug of Vi or 15 ug of Pn6B-TT at 3 wk intervals. The monkeyswere bled before each injection and 3 wk after the last injection .

was elicited by Vi-CTxI , adsorbed ; this preparation elicited the lowest levels ofantibodies after the first injection and there were no booster responses after thesecond or third injections . There were differences in the immunogenicity ofconjugates prepared with the same components . The first injection of Vi-CTvIIIand Vi-CTxI , induced an ^-30-fold increase in CT antibodies, which also rosefourfold after the second and third injections (data not shown) . Vi-CTxI , elicitedhigher levels of CT antibodies, as well as Vi antibodies, than Vi-CTv 1II .Juvenile Rhesus.

A single injection of Vi elicited Vi antibodies in five of thesix monkeys (p = 0 .004) (Table IV). The levels of Vi antibodies declined in allsix monkeys after the second injection of the Vi . Seven of eight monkeys injectedwith Vi-CTxI, responded with a ^-20-fold increase in antibodies after the firstinjection. The one nonresponding monkey had a preimmune level of 0.35 ugantibody per milliliter . The second injection of the conjugate in this groupelicited about a threefold increase over that induced by the first injection of Vi-CTxI , (p = 0.02) and -60 times the preimmunization level (p = 0.0003) . Themonkey that did not respond to the first injection of Vi-CTxI , had a twofoldincrease in antibodies after the second injection . No change in Vi antibodies wasobserved after injection of the Pn6B-TT (control) .

The pathogenic and protective roles of the Vi in typhoid fever have beencontroversial (1, 2, 4, 7, 41-43) . There are data that the Vi exerts the same"shielding" effect upon S. typhi as do the CPS of the encapsulated bacterialpathogens (2, 44-47) . Now there is clinical evidence that antibodies elicited byVi confer immunity against typhoid fever (5, 6) . The pathogenic and protectiveroles of the Vi are therefore, similar to the CPS of other encapsulated bacterialpathogens (47, 48).The main, ifnot the sole, protective immune response elicited by CPS is serum

antibodies . Both the seroconversion rates and postimmunization levels of anti-bodies induced by CPS have been correlated with their protective actions . Serumantibody responses induced by Vi in habitants of areas where typhoid fever isendemic, where there is also a high rate of malnutrition and other acute andchronic infections diseases, are less than optimal (2, 5, 10-12) . Accordingly, we

Antibody*Vaccine n

Pre-immune 1st injection 2nd injection,ug/ml Ag/ml jug/ml

Vi 6 0 .07 (0.05-0.10) 0.23$ (0.06-0.82) 0.07§ (0.05-0.09)Vi-CTX� 8 0 .07° (0.03-0.19) 0.93' (0.24-3.55) 3.65** (0.72-18 .5)Pn6B-TT 8 0.08 (0.06-0.11) 0.10 (0.05-0.21) 0.10 (0.04-0.23)

SZU ET AL .

151 9

covalently bound the Vi to T-dependent proteins both to increase its immuno-genicity and to confer upon it the properties of T-dependency (19, 21, 49). Thissynthesis offers several advantages, namely: (a) there is no crosslinking of eithercomponent; (b) the reactions are conducted in aqueous solutions at neutral pH;(c) the synthesis is applicable to other polysaccharides with carboxyl functions.We have synthesized protein conjugates of Escherichia coli K93, pneumococcustype 12 and Staphylococcus aureus type 8 by this scheme and are studying theimmunological properties of these conjugates (Fattom, A., W. Vann, S. C. Szu,R. Schneerson, et al .) ; and (d) the carrier protein is only slightly modified andcan elicit antibodies to the native protein (25) . The resultant conjugates elicitedhigher levels of antibodies than the Vi alone and induced a booster response inweanling mice andjuvenile primates . Similar studies of Haemophilus influenzaetype b-protein conjugates were predictive of their enhanced antibody responsesin humans (15-17, 19-22) . Whether Vi conjugates will elicit a higher serocon-version rate and levels of antibodies in habitants of areas with high rates oftyphoid fever must be ascertained by clinical evaluation .IgG Vi antibodies protected mice challenged with S. typhi (41, 50, 51). Indirect

evidence was provided that the Vi elicited IgG antibodies in humans (50) . H.influenzae type b-protein conjugates elicit IgG CPS antibodies (19-22). It islikely therefore, that Vi conjugates will also elicit IgG antibodies . We plan toevaluate the isotypes, IgG subclasses and the isoelectric pattern of Vi antibodieselicited by Vi conjugates .

Convalescence from typhoid fever does not always confer immunity to S. typhi(52, 53). Recent data may provide an explanation for this finding; convalescencefrom typhoid fever does not always result in an elevation of Vi antibodies (31,54-56) . Tsang et al ., reported that the Vi alone was a better immunogen in micethan S. typhi strain 560Ty (56) . Increasing evidence points to superior immuno-genicity of CPS-protein conjugates compared with that of the homologousbacteria in certain circumstances, namely : (a) conjugates of H. influenzae type belicit higher levels of CPS antibodies in infants and young children than dosystemic infections (15, 21); and (b) one to three injections of these conjugateselicited higher levels of antibodies in mice than that elicited by the homologousbacteria (Schneerson, R., andJ. B. Robbins, unpublished observations) .

Unlike the CPS of many encapsulated bacterial pathogens, Vi induced serumantibodies and conferred protection against lethal challenge with S. typhi in mice(8). Heidelberger et al . (49) showed that reinjection of several types of pneumo-coccal CPS did not induce a booster response . Landy reported similar data withthe Vi in mice and humans (8, 42). We confirmed that reinjection of Vi did notinduce a booster effect in mice . In Rhesus monkeys, reinjection reduced thelevels of Vi antibodies . This could explain the findings of Gaines et al . (57) whoinjected chimpanzees three times with Vi . The immunized chimpanzees had nodetectable Vi antibodies and were not protected against challenge with S. typhi.The serum antibody responses elicited by conjugates prepared with SPDP

were higher than those prepared by direct binding with EDAC. An explanationfor this finding is that crosslinking reagents, such as SPDP, form conjugates witha "spacer" between the two macromolecules . This property may enable a more

1520


effective interaction between the carrier protein and helper T lymphocytes (15,18, 27, 58).The higher levels of Vi antibodies elicited by the Vi-CT conjugates could be

explained by the adjuvant effect exerted by the residual activity (toxicity) of theCT (20, 60, 61). The formation of a conjugate with the Vi reduced the toxicityof the CT _1O3-104-fold.The resultant conjugate was lethal for the guinea pig,but not mice, in the general safety test for bacterial vaccines described in theCode of Federal Regulations. This was probably due to the greater sensitivity ofguinea pigs to the lethal effects of cholera toxin. We plan to conjugate the Vi tothe ,Q subunit of CT to avoid the problem of toxicity and yet induce serumantibodies that may exert protective effects against other enterotoxigenic bacte-rial pathogens (62) .Some of the primates in these experiments and healthy individuals in the

United States had preexisting serum Vi antibodies (19) . Since S. typhi or S.paratyphi C are rare in the United States, it is improbable that either of thesetwo pathogens were the stimulus for these Vi antibodies . One possible sourcefor the stimulus for these crossreacting antibodies could be the CPS of severalstrains of S. aureus that have an aminogalacturonic acid moiety in their repeatunit (62) . The higher levels of Vi antibodies in young adults in countries wheretyphoid fever is endemic are likely to have been stimulated by infection with S.typhi .

S . typhi is an inhabitant of and a pathogen for humans only ; there are yet noanimal models or in vitro correlates of immunity that could be used to predictthe protective activities of Vi-conjugates . Only the mucin-enhanced lethal infec-tion of mice has been shown to correlate with the clinical effectiveness of cellulartyphoid vaccines (7). The mouse protection model has been shown to be ameasure of Vi antibodies, whether actively induced or passively administered(42) . We have chosen to measure the Vi antibody response induced by theconjugates in laboratory animals by RIA rather than this bioassay (9). Accord-ingly, we plan to standardize the Vi conjugates by physicochemical and bioassaysas has been done for other conjugates, and to evaluate their clinical propertiesofsafety, immunogenicity, and then their protective actions against typhoid feverwithout the use of animal models (22) .

SummaryThe Vi has proven to be a protective antigen in two double masked, controlled

clinical trials in areas with high rates of typhoid fever (-I% per annum) . In bothstudies the protective efficacy of the Vi was ^-70% . ^-75% of subjects in theseareas responded with a fourfold or greater rise of serum Vi antibodies. Incontrast, the Vi elicited a fourfold or greater rise in 95-100% of young adultsin France and the United States . Methods were devised, therefore, to synthesizeVi-protein conjugates in order to both enhance the antibody response and conferT-dependent properties to the Vi (and theoretically increase its protective actionin populations at high risk for typhoid fever) . We settled on a method that usedthe heterobifunctional crosslinking reagent, N-succinimidyl-3-(2-pyridyldithio)-propionate (SPDP), to bind thiol derivatives of the Vi to proteins . This syntheticscheme was reproducible, provided high yields of Vi-protein conjugates, and was

applicable to several medically relevant proteins such as diphtheria and tetanustoxoids . The resultant conjugates were more immunogenic in mice and juvenileRhesus monkeys than the Vi alone . In contrast to the T-independent propertiesof the Vi, conjugates of this polysaccharide with several medically relevantproteins induced booster responses in mice and in juvenile Rhesus monkeys .Clinical studies with Vi-protein conjugates are planned . This scheme is alsoapplicable to synthesize protein conjugates with other polysaccharides that havecarboxyl functions .

Helpful discussions with Dr . John Inman are gratefully acknowledged. Dr . William Eganassisted us with the NMR analysis ofthe Vi antigens . Dr .Joseph Shiloach, Pilot ProductionPlant, LOMB, NIDDK, NIH cultivated the C. freundii and started the purification of theVi . Ms . Dolores M. Bryla performed the statistical analyses. The expert technical assistanceof Mr . Tod Cramton is gratefully acknowledged . Drs. Margaret Pittman and Charles U.Lowe made helpful suggestions and comments in their review of this manuscript . Dr .Dominque Schulz and Dr . Jacques Armand, Institut Merieux, Lyon, France, kindlyprovided us with technical advice, tetanus toxoid, cholera toxin, and Vi CPS . Dr . RudiTiejsema and Coen Beuvery, Rijks Instituut voor Volksgezonheid, Bilthoven, Nether-lands, kindly provided us with diphtheria toxoid .

Received for publication 16July 1987.

SZU ET AL.

152 1

References1 . Edelman, R., and M. M. Levine . 1986 . Summary of an international workshop on

typhoid fever . Rev. Infect. Dis. 8 :329 .2 . Hornick, R. B . 1985 . Selective primary health care : strategies for control of disease

in the developing world . XX. Typhoid fever . Rev . Infect. Dis. 7:536 .3 . Hornick, R . B ., and T. E . Woodward. 1966 . Appraisal of typhoid vaccine in experi-

mentally infected human subjects . Trans . Am. Clin . Climatol. Assoc . 78 :70.4 . Levine, M . M., C . Ferreccio, R . E . Black, R . Germanier, and Chilean Typhoid

Committee . 1987 . Large-scale field trial of Ty21 a live oral typhoid vaccine in enteric-coated capsule formulation . Lancet. i :1049 .

5 . Acharya, I . L ., R . Tapa, V . L . Gurubacharya, M. B . Shrestha, C . U . Lowe, D . A .Bryla, R . Schneerson, J . B . Robbins, T. Cramton, B . Trollfors, M. Cadoz, D . Schulz,and J . Armand . 1987 . Prevention of typhoid fever in Nepal with the Vi capsularpolysaccharide of Salmonella typhi : A preliminary report. N. Engl. J. Med. I n press.

6 . Klugman, K. P ., I . Gilbertson, H. J . Koornhof, D . Schulz, M. Cadoz, J . Armand, andJ . B . Robbins . 1987 . The prevention of typhoid fever by capsular polysaccharide(CPS) vaccination . ICAAC, 27th, New York. (Abstr .) .

7 . Robbins, J . D., and J . B . Robbins . 1984 . Re-examination of the immunopathogenicrole of the capsular polysaccharide (Vi antigen) of Salmonella typhi. J. Infect. Dis.47 :49.

8 . Landy, M. 1954 . Studies in Vi antigen . VI . Immunization of human beings withpurified Vi antigen . Am. J. Hyg. 60:52 .

9 . Tackett, C . O ., C . Ferriccio, J . B . Robbins, C-M . Tsai, D . Schulz, M . Cadoz, A.Goodeau, and M. M. Levine . 1986 . Safety and characterization of the immuneresponse to Salmonella typhi Vi capsular polysaccharide vaccines . J. Infect. Dis.154:342 .

10 . Greenwood, B . M . 1984 . Selective primary health care : strategies for control ofdisease in the developing world. XIII . Acute bacterial meningitis . Rev. Infect. Dis.6(Suppl .):S374 .

1522


11 . Reingold, A . L ., C . V . Broome, A . W. Hightower, G . W. Ajello, G . A . Bolan, C .Adamsbaum, E . E . Jones, C . Philips, H . Tiendrebeogo, and A. Yada . 1985 . Menin-gococcal polysaccharide A vaccine : evidence ofage-specific differences in the durationof clinical protection following vaccination . Lancet . ii : 114 .

12 . Williamson, W. A., and B . M . Greenwood . 1978 . Impairment of the immune responseto vaccination after acute malaria . Lancet. i :1328 .

13 . Avery, O . T., and W. F . Goebel . 1929 . Chemo-immunologica l studies on conjugatedcarbohydrate-proteins . 11 . Immunological specificity of synthetic sugar-proteins . J.Exp. Med. 50:521 .

14 . Goebel, W. F . 1929 . Chemo-immunological studies on conjugated carbohydrate-proteins . XII . The immunological properties of an artificial antigen containingcellobiuronic acid . J. Exp. Med. 50:469 .

15 . Anderson, P . W., M . E . Pichichero, R . A . Insel, R . Betts, R . Eby, and D . H . Smith .1986 . Vaccine s consisting of periodate-cleaved oligosaccharides from the capsule ofHaemophilus infuenzae type b coupled to a protein carrier : structural and temporalrequirements for priming in the human infant .J. Immunol. 137:1181 .

16 . Beuvery, E . C ., F . V . Rossum, andJ . Nagel . 1982 . Comparison ofthe and G antibodiesin mice with purified pneumococcal type 3 and meningococcal group C polysaccha-rides and their protein conjugates . Infect. Immun. 40:245 .

17 . Chu, C . Y ., R . Schneerson, J . B . Robbins, and S . C . Rastogi . 1983 . Further studieson the immunogenicity of Haemophilus infuenzae type b and pneumococcal type 6Apolysaccharide-protein conjugates . Infect . Immun. 40:245 .

18 . Jorbeck, H . J . A ., S . B . Svenson, and A . A . Lindberg . 1981 . Artificia l Salmonellatyphimurium 0-antigen-specific oligosaccharide-protein conjugates elicit opsoninizingantibodies that enhance phagocytosis . Infect. Immun. 32:497 .

19 . Schneerson, R., O. Barrera, A . Sutton, and J . B . Robbins. 1980 . Preparation,characterization and immunogenicity ofHaemophilus infuenzae type b polysaccharide-protein conjugates . J. Exp. Med. 152 :361 .

20 . Schneerson, R., J . B . Robbins, C.-Y . Chu, A. Sutton, W. Vann, J . C . Vickers, W. T .London, B . Curfman, M. C . Hardegree, J . Shiloach, and S . C . Rastogi. 1984 . Serumantibody responses ofjuvenile and infant Rhesus monkeys injected with Haemophilusinfuenzae type b and pneumococcus type 6A polysaccharide-protein conjugates .Infect. Immun. 45:582 .

21 . Schneerson, R., J . B . Robbins, S . C . Szu, and Y . Yang . 1987 . Vaccines composed ofpolysaccharide-protein conjugates : current status, unanswered questions, and pros-pects for the future. In Towards Better Carbohydrate Vaccines . R . Bell and G.Torrigiani, editors . John Wiley & Sons Ltd, Chichester, United Kingdom . In press .

22 . Schneerson, R., J . B . Robbins,J . C . Parke, Jr ., A . Sutton, Z . Wang, J . J . Schlesselman,G. Schiffman, C . Bell, A . Karpas, and M. C . Hardegree . 1986 . Quantitative andqualitative analyses of serum Haemophilus infuenzae type b, pneumococcus type 6Aand tetanus toxin antibodies elicited by polysaccharide-protein conjugates in adultvolunteers. Infect. Immun. 52:501 .

23 . Szu, S . C ., R . Schneerson, and J . B . Robbins. 1986 . Rabbit antibodies to the cell wallpolysaccharide of Streptococcus pneumoniae fail to protect mice from lethal infectionwith encapsulated pneumococci . Infect. Immun. 54:448 .

24 . Heyns, K., and G. Kiessling . 1967 . Strukturafklarung des Vi-antigens aus Citrobacterfreundii (E . coli) 5396/38 . Carbohydr. Res. 3:340 .

25 . Carlsson, J ., H . Drevin, and R . Axen. 1978 . Protein thiolation and reversible protein-protein conjugation . Eur. J. Biochem. 173:723 .

26 . Anderson, P., J . Pitt, and D . H . Smith . 1976 . Synthesi s and release of polyribophos-phate by Haemophilus infuenzae type b in vitro . Infect. Immun. 13:581 .

SZU ET AL .

152 3

27 . Katz, M., and A. M . Pappenheimer, Jr. 1969 . Quantitative studies of the specificityof anti-pneumococcal antibodies, types III and VIII . IV . Binding of labeled hexasac-charides derived from S3 by anti-S3 antibodies and their fab fragments . J. Immunol.103:491 .

28 . Gotschlich, E . C ., M . Rey, W. R . Sanborn, R. Triau, and B . Cvjetanovic . 1972 . Theimmunological responses observed in field studies in Africa with Group A meningo-coccal vaccines. Prog. Immunobiol . Stand. 129 :485 .

29 . Tsai, C-M . 1986 . The analysis of lipopolysaccharide (endotoxin) in meningococcalpolysaccharide vaccines by silver staining following SDS-polyacrylamide gel electro-phoresis . J. Biol . Stand. 14:25 .

30 . Szu, S . C ., G . Zon, R . Schneerson, and J . B . Robbins . 1986 . Characterization of thedepolymerization of bacterial polysaccharides induced by ultrasonic irradiation . Car-bohydr . Res. 152:7 .

31 . Craig, J . P . 1965 . A permeability factor (toxin) found in cholera stools and culturefiltrates and its neutralization by convalescent cholera sera . Nature (Lond.) . 207:614 .

32 . Fowler, B., and A. J . Robbins. 1972 . Methods for the quantitative analysis of sulphur-containing compounds in physiologic fluids . J. Chromatogr . 72 :105 .

33 . Ellman, G . I . 1959 . Tissue sulfhydryl groups . Arch. Biochem. Biophys. 82:70 .34 . Bradford, M. M . 1976 . A rapid and sensitive method for the quantitation of micro-

gram quantities of protein utilizing the principle of protein-dye binding . Anal .Biochem. 72:248 .

35 . Stone, A . L ., and S . C . Szu . 1987 . The optical properties of the Vi-capsular polysac-charide : application in analysis of the Vi in vaccines . J. Clin . Microbiol. In press .

36 . Elson, C . O., and W. Ealding . 1985 . Genetic control of the murine immune responseto cholera toxin . J. Immunol. 135:930 .

37 . Nolan, C . M .,J . C . Feeley, P . C . White, E . A . Hambie, S . L . Brown, and K.-H . Wong.1980 . Evaluation of a new assay for Vi antibody in chronic carriers of Salmonellatyphij Clin . Microbiol. 12:22 .

38 . Dafni, Z ., and J . B . Robbins . 1976 . Purification of heat-labile enterotoxin fromEscherichia coli 078:H I I by affinity chromatography with antiserum to Vibrio choleratoxin .J. Infect. Dis. 133(Suppl .):S138 .

39 . Whiteside, R . E ., and E . E . Baker . 1961 . The Vi antigens of the enterobacteriaceae .J. Immunol. 86:538 .

40 . Wong, K. H., and J . Feeley . 1972 . Isolation of Vi antigen and a simple method forits measurement . Appl. Microbiol. 24:628 .

41 . Gaines, S .,J . A . Currie, and J . G . Tully . 1960 . Production of incomplete Vi antibodyin mice . Proc . Soc. Exp. Biol. Med. 104:602 .

42 . Landy, M. 1957 . Studies on the Vi antigen . VII . Characteristics of the immuneresponse in the mouse . Am. J. Hyg. 65:83 .

43 . Felix, A., and R . M . Pitt . 1953 . The pathogenic and immunogenic activities ofSalmonella typhi in relation to its antigenic constituents . J. Hyg. 49:92 .

44 . Kauffmann, F . 1936 . Untersuchunger uber die Korperantigene in der Salmonella-Gruppe . Z. Hyg. Infektionskr. 117:778 .

45 . Kauffman, F . 1954 . Enterbacteriaceae . Munksgaard, Copenhagen. 50 .46 . Kenne, L., and B . Lindberg . 1983 . Bacterial polysaccharides . In The Polysaccharides .

Vol . 2 . G . O . Aspinall, editor . Academic Press, New York . 315 .47 . Robbins, J . B . 1978 . Vaccines for the prevention of encapsulated bacterial diseases :

current status, problems and prospects for the future . Immunochemistry. 15:839 .48 . Robbins, J . B ., R . Schneerson, W. B. Egan, W. Vann, and D . T. Liu . 1980 . Virulence

properties of bacterial capsular polysaccharides . Unanswered questions . In The

1524


Molecular Basis of Microbial Pathogenicity . H . Smith, J . J . Skehel, and M. J . Turner,editors . Verlag Chemie GmbH, Weinheim . 115-132 .

49 . Heidelberger, M. 1953 . Persistance of antibodies in man after immunization . InNature and Significance of Antibody Response . A . M . Pappenheimer, Jr ., editor .Columbia University Press, New York .

50 . Gaines, S .,J . A . Currie, andJ . G . Tully . 1965 . Production ofincomplete Vi antibodyin man by typhoid vaccine . Am. J. Epidemiol. 81 :350 .

51 . Kawata, Y . 1970 . A study of the molecular types of immunoglobulin . II . Mouseprotection study of Vi antibody against typhoid infection . Acta Medicine Univ. Kioto.40:284 .

52 . Islam, A., T . Butler, and L . R . Ward. 1987 . Reinfection with a different Vi-phagetype ofSalmonella typhi in an endemic area . J . Infect. Dis. 155:155 .

53 . Marmion, D . E ., G . R . E . Naylor, and I . O . Stewart . 1953 . Second attacks of typhoidfever . J. Hyg. 51 :260.

54 . Brodie, J . 1977 . Antibodies and the Aberdeen typhoid outbreak of 1964 . 1 . TheWidal reaction . J. Hyg. Cambridge. 79:161 .

55 . Lanata, C . F ., C . Ristori, L . Jimenez, G. Garcia, M . M . Levine, R . E . Black, M .Salcedo, and V . Sotomayor. 1983 . Vi serology in detection of chronic Salmonellatyphi carriers in an endemic area. Lancet . i :441 .

56 . Tsang, R . S . W., and P . Y. Chau . 1987 . Productio n of Vi monoclonal antibodies andtheir application as diagnostic reagents . J. Clin. Microbial. 25:531 .

57 . Gaines, S ., M . Landy, G. Edsall, A . D . Mandel, R-J . Trapani, and A . S . Beneson .1961 . Studies on infection and immunity in typhoid fever.J. Exp. Med. 114:327 .

58 . Fong, S., D . E . Nitecki, R . M . Cook, andJ . W . Goodman . 1978 . Spatia l requirementsbetween haptenic and carrier determinants for T-dependent antibody responses . J.Exp. Med. 148:817 .

59 . Toshinori, K., C-Y Lawn, A. Amsden, and S . Leskowitz . 1983 . Hapten-specific T cellresponse to azobenzenearsonate-N-acetyl-L-tyrosine in the Lewis rat . III . Effects ofpeptide-spacer structure on eliciting ASA-specific helper activity with TNP-haptenedABA-peptide-Ficoll . J. Immunol. 130:586 .

60 . Chisari, F . V ., R . S . Northrup, and L . C . Chen. 1974 . The modulating effect ofcholera enterotoxin on the immune response . J. Immunol. 113:729 .

61 . Finkelstein, R . 1984 . Cholera . In Bacterial Vaccines . R . Germanier, editor . AcademicPress, New York . 107-136 .

62 . Liau, D-F ., A . Melly, andJ . H . Hash. 1974. Surface polysaccharide from Staphylococcusaureus M that contains taurine, D-aminogalacturonic acid and D-fucosamine . Infect.Immun. 119:913 .

vi capsular polysaccharide-protein conjugates · pdf filevi capsular polysaccharide-protein...

Documents